C.W. Ndung’u*, T. K. Muasya, and T. O. Okeno
Animal Breeding and Genomics Group, Department of Animal Sciences,
Egerton University, P. O. Box 536-20115, Egerton, Kenya
*Corresponding Author: firstname.lastname@example.org
This study hypothesized that use of genetic parameters from different studies to evaluate overall genetic and economic gains of livestock breeding programmes could over and under estimate response to selection. This is because genetic parameters are affected by data sample size, environmental conditions and evaluation models. This premise was tested by deterministic simulation of breeding schemes that resemble that used in indigenous chicken in Kenya. Rates of genetic gain and inbreeding were estimated when non-pooled (NPP) and pooled (PP) genetic parameters were adopted. The NPP represented the conventional evaluation model where genetic parameters for traits are sourced from different studies assuming factors affecting their predictions. The PP strategy mimicked an alternative model cognizant of the factors affecting the parameters and therefore parameters from different studies were subjected to meta-analysis to obtain consensus estimates. Response to selection after one round of selection in the two strategies was compared assuming a single tier nucleus breeding scheme with 480 breeding candidates with mating ratio of 1:5 for cocks and hens. The rates of genetic gain and inbreeding were KES 40.50 and 0.68% and KES 61.10 and 0.49% for PP and NPP, respectively. Rate of genetic gain and inbreeding realized in PP was 1.5 times lower and 38% higher than that realized in NPP, respectively. The findings in this study confirm the premise that, accounting for variations through meta-analysis in estimation of input genetic parameters minimizes the over and under estimation of genetic gain and inbreeding.
Keywords: Indigenous chicken, Meta-analysis, Response to selection